Model-data fit evaluation: Aberrant response detection

Abstract
Person-fit analysis is an important field aimed at establishing the validity (or lack thereof) of the response pattern provided by each respondent of a test or questionnaire. This chapter discusses person-fit analysis within the framework of Item Response Theory. We illustrate how aberrant response patterns can look like and what are the most popular statistical strategies currently in used to detect them. Examples of empirical applications of these techniques are also provided. Our goal is to motivate practitioners to this important quality-check aspect of test scoring.
Type
Publication
International Encyclopedia of Education 4e